Skip to main content

Full text of "Mechanisms for the mutarotation and hydrolysis of the glycosylamines and the mutarotation of the sugars"

See other formats


Journal of Research of the National Bureau of Standards 



Vol. 46 No. 2, February 1951 



Research Paper 2186 



Mechanisms for the Mutarotation and Hydrolysis of the 
Glycosylamines and the Mutarotation of the Sugars 



Horace S. Isbell and Harriet L. Frush 



A study has been made of the kinetics of the mutarotation and hydrolysis reactions of 
L-arabinosylamine, and a mechanism has been devised to account for the striking sensitivity 
of the glycosylamines to hydrolysis in a limited pH range. The concepts presented seem 
applicable for the interpretation of the reactions of other compounds of the aldehyde 
ammonia type. 



I. Introduction 

In the development of methods for the preparation 
of the amides of uronic acids [1] * it became necessary 
to determine conditions whereby the amino group 
of a 1-amino-uronic amide could be hydrolyzed 
without alteration of the amide group. It was 
found that the rate of hydrolysis of the amino group 
of 1-aminomannuronic amide (I) is extraordinarily 
sensitive to the acidity of the reaction medium. 
Thus, when the compound is mixed quickly with 
one equivalent of a strong acid, hydrolysis requires 
a period of many hours, but when the acid is added 
dropwise to the compound in solution, hydrolysis is 
complete in 15 min or less. Further study showed 
that 1-aminomannuronic amide is surprisingly 
stable both to strong acid and to alkali, and that 
hydrolysis can be effected rapidly only in the pH 
range 4 to 7. Because of the presence of the hy- 
drolyzable amide group in 1-aminomannuronic 
amide, the study of the hydrolysis was directed to 
the 1-aminosugars, more properly designated gly- 
cosylamines. 

Exploratory experiments with glucosylamine, 
galactosylamine, and arabinosylamine showed that 
the glycosylamines in general possess the peculiar 
sensitivity to hydrolysis in a limited pH range that- 
had been noted for 1-aminomannuronic amide [2]. 
In this paper, the mutarotation and hydrolysis of 
L-arabinosylamine are considered in detail. Mecha- 
nisms for the mutarotation of the sugars, comparable 
to those for the mutarotation of the glycosylamines, 
correlate the properties of the two groups of sub- 
stances and account for the over-all rate of mutaro- 
tation of both the glycosylamines and the sugars as 
a function of pH. 



0=C C 

I 
NH 2 



H H OH OH 

I I I I 

C C C — CH-NHa 

I I I 

OH H H 

O— 



1-Aminomannurouic amide 



1 Figures in brackets indicate the literature references at the end of this paper. 



OH OH H 

I I I 
H 2 C— C C C CH-NH 2 

I I I 

H H OH 



-0- 



II. 



L- Arabinosylamine 

OAc OAc H 

I I I 
H 2 C— C C C CH-NH-Ac 





1 1 1 
H H OAc 






o 










Tetraacet yl-L-arabinosy lamine 


OH OH H 

1 1 


H 2 C— C C C CHNHAc 




1 1 1 
H H OH 





III. 



IV. 



N- Acetvl-L- arabinosvlamine 



II. Structure and Chemical Properties of 
L-Arabinosylamine 

L-arabinosylamine (II) was prepared by treating 
L-arabinose in methanol with ammonia essentially by 
the method of de Bruyn and Van Leent [3]. Acetyla- 
tion of the compound with acetic anhydride and 
pyridine yielded a new tetraacetate (III). Catalytic 
deacetylation of this substance with barium methy- 
late in methanol gave a crystalline A T -acetyl-L- 
arabinosylamine. 2 The latter substance was found 
to react quickly with 2 moles of sodium periodate, as 
required for an iV-acetylpentosylamine having the 
pyranose structure IV. Since acetylation and de- 
acetylation were carried out by mild reactions that 
ordinarily cause no change in ring structure, the 
original L-arabinosylamine and its tetraacetate are 
tentatively classified as pyranoses. 

In a slow titration of L-arabinosylamine, one 
equivalent of acid was required for neutralization 

2 This compound is analogous to the TVacetyl-D-glucosy lamine prepared by 
Brigl and Keppler [4] and shown by Niemann and Hays [5] to be a pyranose. 



132 



8 r 

7 

6 

4 



0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 I.I 1.2 1.3 
EQUIVALENTS HCL/ MOLE L-ARABINOSYL AMINE 

Figure 1. Titration of L-arabinosylamine with acid. 

(fig. 1). Separation of crystalline L-arabinose in 
nearly quantitative yield from the mixture after 
titration established the fact that this treatment with 
dilute acid had caused hydrolysis of the glycosyla- 
mine and formation of the free sugar. 

When dissolved in water, L-arabinosylamine under- 
goes spontaneous changes that give rise to a decrease 
in optical rotation and then to an increase. The 
changes can be ascribed principally to a mutarota- 
tion reaction that establishes equilibrium between 
the various modifications of the glycosylamine, and 
hydrolysis of the amino group. 3 The relative rates 
of the two reactions vary with experimental condi- 
tions (fig. 2). Thus, when L-arabinosylamine is 
dissolved in strong acid, the optical rotation drops 
almost at once to about +69°, and then increases 
slowly over a period of several weeks (curve I). 
However, if it is dissolved in weak acid, for example 
in water saturated with carbon dioxide, the optical 
rotation drops immediately to about. +70° and then 
increases rapidly to +105° (curve II). In a weakly 
alkaline solution prepared by dissolving the sub- 
stance in carbon dioxide-free water, the specific 
rotation decreases in the course of a few hours from 
+86° to a minimum, and then slowly rises (curve 
III). In strong alkali there is almost no change 
(curve IV). 

These qualitative observations may be summa- 
rized as follows: 



Solution 



Strongly acid_ 
Weaklv acid__ 



Weakly alkaline. _ 
Si rongly alkaline 



Mutarotation 
reaction 



Very rapid _ 
__do 



Rapid, but meas- 
urable. 
Slow 



Hydrolysis reaction 



Slow. 

Rapid, but meas- 
urable 
Slow. 

Do. 



To obtain quantitative data as to the effect of the 
oxoninm ion concentration, samples of L-arabinosyl- 



3 The process appears to be complicated under some conditions by the formation 
of the diglyeosylamine, and possibly by other reactions that depend on the 
presence of an Intermediate imine cation. See page 142. 



amine were treated with acids and bases in solutions 
buffered at various pH values, and the ensuing 
reactions were followed by optical rotation. Al- 
though the mutarotation and hydrolysis reactions of 
L-arabinosylamine take place simultaneously, they 
can be studied separately because their rates at a 
given pH value differ widely, and because the change 
in optical rotation accompanying each of the two 
reactions is large. For study, the change in optical 
rotation was considered to consist of two periods: a 
short period beginning at zero time and character- 
ized by a decrease in dextrorotation; and a long 
period beginning when the initial change was com- 
plete, characterized by an increase in dextrorotation 
and extending until the optical rotation reached a 
value corresponding to complete hydrolysis (+105.2°). 
The rate of the mutarotation reaction was obtained 
from the data for the first period and the rate of the 
hydrolysis reaction from the data for the second 
period. Suitable data were obtained for the mutaro- 
tation reaction in alkaline solutions, in which the 
hydrolysis is extremely slow, and for the hydrolysis 
reaction in weakly alkaline or acid solutions, in which 
the mutarotation reaction is almost instantaneous. 

Satisfactory rate constants for the mutarotation of 
L-arabinosylamine were obtained in the pH range 7.8 
to 12 by application of the customary formula 4 for 
a first-order reaction to the data of table 1 . Table 2 
gives the mutarotation constants that were obtained 
for a series of experiments used to evaluate the 
catalytic effects of the oxonium, hydroxyl, and am- 
monium ions. The dotted curve of figure 3 represents 
these mutarotation constants corrected for the cata- 
lytic effect of tin 4 ammonium ion. The results 
clearly show that the mutarotation of L-arabino- 
sylamine is strongly catalyzed by acids but not ap- 
preciably by bases. 5 The mutarotation differs from 



+ 110 

105 

100 

| 95 

o 90 

g 85 

o 

£ 80 

CD 

75 

70 

65 

4 8 12 16 20 24 28 32 36 40 44 48 52 56 
TIME , HOURS 

Figure 2. Mutarotation of L-arabionsylamine. 

I, In 2.5 ATHC1; II, in C 2 -saturated water; III, in C0 2 -free water; IV, in 0.01 
iVNaOH. 













s 




















< 


y^ 


n 




















































































ez: 








— c 


\ 






































«o= 












IE 


-o- 
















-O— — 






i 


















— o 


1 





























<The first-order formula is applicable if the concentrations of the acid and base 
catalysts are held constant. 

5 However, unpublished work has shown that the mutarotation of a-D-galacto 
sylamine is weakly but definitely catalyzed by bases. 



133 



the mutarotations of the sugars in that the latter are 
catalyzed by bases more strongly than by acids. The 
reason for this difference will be considered in the 
next section. 

Table 1. Mutarotation and hydrolysis measurements 

0.5 g of L-arabinosylamine dissolved in sufficient acid, base, or buffer to give a 
volume of 25 ml. Solvents are listed opposite experiment numbers. 



Time 
(minutes) 



10 
32 
60 
120 
270 
1,250 



1.7 

15 

60 

125 

300 

1,290 



1.6 
6.5 
8.7 
17.0 
30.2 



l.« 

3.8 
4.8 
6.8 
21.fi 



-IS 



pH 



Mutarotation con- 
stant^ 
1 , r h-r™ 
h—U r i2 —r m 



Hydrolysis rate 
constant, a A'hydroi 

■; — 7 log 

t2-ti r t0 -r» 



Experiment 1. 0.01 IV NaOH 



+86.4 




86.1 




86.0 




85.1 




82.7 


12.1 


77.6 




b (69.0) 


Avg._ 



0. 00034 
. 00040 
. 00031 
. 00040 



0. 00036 



Experiment 2. 0.1 N HC1+NH 3 to pH 10.5 



+85.0 




84.2 




82.2 




79.6 


10.5 


75.5 




70.3 




b (69.0) 






Avg.. 



0.0017 
. 0011 
.0015 
.0013 



Experiment 3. 2.5 N HCI+NH3 to pH 10.5 



3.3 


+78.7 
77.2 
73.1 
67.7 
65.3 

* (61.9) 


10.5" 

Avg_. 






4.9 


0.025 
.026 
.027 
.026 




10.2 




20.1 




30.2 




120.0 












0.026 





Experiment 4. 0. 1 N HCI+NH3 to pH 8.0 



2.2 


+82.2 
79.2 
76.0 
73.2 
66.9 


"~8.Y 
Avg__ 






6.1 


0.024 
.023 
.019 




12. 1 




21.1 




145 












0.022 





Experiment 5. 1 N HCI+NPI.3 to pH 8.0 



+81.5 
75.6 
74.0 
70.7 
69.0 



Avg. 



0.057 
.056 
.056 



0.056 



Experiment 6. 1 N HCI+NH3 to pH 7.5 



+78.2 
74.5 
73.2 
71.3 
68.9 



Avg- 



0.116 
.116 
.118 



Table 1. Mutarotation and hydrolysis measurements — Con. 

0.5 g of L-arabinosylamine dissolved in sufficient acid, base, or buffer to give a 
volume of 25 ml. Solvents are listed opposite experiment numbers. 



Time 
(minutes) 


[*}% 


pH 


Mutarotation con- 
stant, a 
1 , r h -rm 


Hydrolysis rate 

COnStant, a fchydrol 

1 r tl —r x 

=- — 7 log 

t2—U r t2 -rn 


C2—11 r t2 —r m 


Experiment 7. 1 volume of 1 M KH2PO4+I volume of 1 N NaOH, pH 10.0 


2.7 
4.3 
6.0 
12.8 
17.1 
60.0 


+82.4 
81.5 
80.4 
76.4 
74.7 
70.3 

b (69. 0) 

+72.2 

81.6 

84.6 

85.1 

d (105. 2) 


"16.6" 

Avg-- 

16." 4" 






0.019 
.021 
.026 
.025 
















210.0 
1, 750 
3. 570 
4,560 

00 






0.023 






0. 00009 
. 00006 
. 00005 








Avg-- 




0. 00007 


Experiment 8. 1 volume of 1 N Na 2 C0 3 +l volume of 1 iVNaHC0 3 , pH 9.4 


2.0 
3.8 
4.6 
5.5 
10.1 
30.0 


+82.2 
79.2 
78.0 
77.0 
73.6 
70.9 

b (69. 0) 

+74.5 

78.1 

82.6 

84.3 

d (105. 2) 


" _ 9.T 

Avg- 
"~9~5~ 






0.060 
.063 
.061 
.055 
















420 
1,380 
2,800 
4,320 






0.060 








0. 000056 
. 000056 
. 000043 








Avg-- 




0. 000051 


Experiment 9. 1 volume of 1 JV Na 2 CO 3 +10 volumes of 1 N NaHC0 3 , pH 8.6 


1.6 
2.4 
5.5 
6.6 
12.1 

180.0 

1,140 

4,320 

8,640 

00 


+79.0 
77.2 
72.9 
72.0 
70.4 

b (69. 0) 

+72.7 

77.8 

88.4 

93.1 

d (105. 2) 


'~8."6~ 

Avg-- 

8.7 






0.108 
.105 
.105 


















0.106 






0. 000077 
. 000069 
. 000051 








Avg- . 




0. 000066 


Experiment 10. 10 volumes of 1M KH2PO4+7 volumes of 1 N NaOH 
diluted to 20 volumes, pH 7.0 


1.4 

6.0 

10.5 

20.4 

1500. 


+71.3 
78.1 
81.5 
87.6 
103.1 


"~7.~2~ 








0.023 
.018 
.016 








Avg_. 




0.019 


Experiment 11. 2 volumes of 1 M KH2PO4+I volume of 12V NaOH 
diluted to 4 volumes, pH 6.6 


1.3 
6.6 
10.0 
19.9 
110.0 


+73.9 
84.0 
87.4 
93.6 
104.3 


"~6.Y 








0. 033 
.029 
.024 








Avg-- 




0.029 



134 



Table J. Mutarotation and hydrolysis measurements — Con. 

0.5 g of L-arabinosvlamine dissolved in sufficient acid, base, or buffer to give I 
volume of 25 ml. Solvents are listed opposite experiment numbers. 



Time 

(minutes) 


W 2 o° 


pll 


Mutarotation con- 
stant, » 
1 r h -r m 


Hydrolysis rate 
constant, a /ch y droi 
1 . r h~r K 

= log 


Experiment 12. 5 volumes of 1 M KH2PO4+I volume of 1 N NaOH 
diluted to 10 volumes. pH 5.9 


1.4 

6.1 

9.6 

11.3 

60.0 


+77.1 
89.5 
94.2 
96.0 
103.9 


"12" 








0.057 
.054 
.054 








Avg._ 




0.055 


Experiment 13. 10 volumes of 1 M KH2PO4+I volume of 1 N NaOH 
diluted to 20 volumes, pH 5.5 


2.9 

5. 6 

9.7 

60. 


+80.1 
90.6 
97.0 
104.2 


"5.9" 








0. 069 
.068 






Avg.- 




0.068 


Experiment 14. 5 volumes of 1 A T acetic acid+3 volumes of 1 N NaOH 
diluted to 10 volumes, pH 4.7 


1.3 
9.9 
12.0 
13.1 
17.9 


+69.4 
100.1 
101. 6 
102.5 
104.2 


"To" 








0.108 
.105 
. Ill 








Avg.. 




0.108 


Experiment 15. 5 volumes of 1 N acetic acid+2 volumes of 1 N NaOH 
diluted to 10 volumes, pH 4.4 


1.5 
6.2 
9.5 
11.2 
12.9 
21.6 


+70.8 
92.4 
99.8 
101.5 
102.9 
104.4 


~~4.~6~ 
"~4.T 








0.095 
.108 
.110 
. 118 










Avg._ 




0.108 


Experiment 16. 1 volume of 5.14 iV acetic acid+1 volume of 1 iVNaOH, 
pH 3.9 


2.3 
6. 6 
8.2 
10.4 
41.9 


+76.0 
90.4 
93.4 
96.3 
104.7 


To" 








0.070 
.069 
.066 








Avg.. 




0.068 


Experiment 17. 2 volumes of 5.14 N acetic acid+1 volume of 1 iVNaOH, 
pH3.5 


1.9 
4.7 
9.7 
14.5 
150 


+70.7 
80.2 
90.1 
95.3 
104.8 


~3.~6~~ 
~3.Y~ 








0.051 
.047 
.044 








Avg._ 




0.047 


Experiment 18. 10 volumes of 5.14 iV acetic acid +1 volume of 1 iVNaOH, 
pH 2.7 


1.8 

2.6 

9.6 

20.0 

105 


+66.5 
68.0 
78.3 
87.6 
104.2 


Ti " 








0.022 
.021 
.021 








Avg_- 




0.021 













Table 1. Mutarotation and hydrolysis measurements — Con. 

0.5 g of L-arabinosylamine dissolved in sufficient acid, base, or buffer to give 1 
volume of 25 ml. Solvents are listed opposite experiment numbers. 



Time 

(minutes) 


MS 


PH 


Mutarotation con- 
stant, 51 

12— 1\ rt 2 —r m 


Hydrolysis rate 

constant, a /chydroi 

1 r h — r m 

= -+ — 7 lo s ; — t~ 

62 — Cl Tt 2 — Too 


Experiment 19. 1 iV oxalic acid, pH 0.8 


1.4 

5.0 
480 
1,440 
1,890 

00 


e +67. 1 

69.3 
72.1 

77.7 

80.6 

d (105. 2) 


"6~8~~ 












0. 000074 

. 000081 
. 000087 








Avg._ 




0. 000081 


Experiment 20. 2.5iVHCl 


3.8 
30.0 
60.0 
3. 300 
35, 300 


e +69. 

e69.9 

70.0 

71.2 

81.2 

<* (105. 2) 


~0.~2~ 
















0. 0000046 
. 0000047 






Avg_. 








0. 0000047 



i Symbols are defined in section VII, 5. 
> See page 144. d Postulated final value. 



b Postulated equilibrium value 
e Value not used in calculations. 



Table 2. Mutarotation constants of L-arabinosylamine at 

20° C 



Experi- 
ment 


pll 


[IP] 


[OH-] 


[NH+] 


km 


km cor- 
rected «■ 


1 

2 

3 

4 

5 

6 


12. 1 
10.5 
10.5 

8.5 
8.2 
7.8 


7.94X10-1* 
3.16X10 " 

:;. ir>xi0-'i 

3. 10X10-& 
6.31X10-9 
1. 58X10-" 


1.26X10-2 
3. 10X10-* 
3. 16X10-* 
3.16X10- 6 
1.58X10- 6 
C). 31X10- 7 


"0. i" 

2.5 
0. 1 
1.0 
1.0 


0.001 

.0016 

. 026 
. 022 

. 056 
. 117 


0.0004 
. OOOfl 
.001 

.021 
.046 
. 107 



a Mutarotation constant, l; m , was corrected for the catalytic effect of ammo- 
nium ion. See page 139. 

III. Mechanisms for the Mutarotation Reac- 
tions of the Sugars and the Glycosylamines 

A satisfactory mechanism for the mutarotation of 
both the sugars and the glycosylamines must lead to 
an accurate expression for the rate of mutarotation 
as a function of the acid and base catalysts present. 
Years ago Hudson showed that the mutarotation 
constants for the sugars in the presence of strong 
acids and bases can be represented by the following 
empirical equation [6]: 

fc w = fcH 2 o + i H [H + ] + ioH[OH-], (1) 

where ku 2 o, k H , and k n are constants characteristic 
of the particular sugar, and [H + ] and [OH - ] represent 
the concentrations of the oxonium and hydroxyl ions, 
respectively. More recently it has been established 
that the reaction is subject to general acid and base 
catalysis, and consequently the expression has been 
extended to include terms accounting for the cata- 
lytic effect of undissociated acids, anions of weak 
acids and cations of weak bases [7]. Thus eq 1 is 
replaced by eq 2: 

^ = &h 2 o + £h[H+]-T- 

*oh[OH-] . . . +±k i [RA i ] + ±k j [B j ], (2) 



135 



in which HA signifies any hydrogen acid, neutral or 
ionic, and B signifies any base, neutral or ionic. 

Although several mechanisms have been advanced 
for the mutarotation reaction, no single mechanism 
accounts for all of the experimental facts. To 
account for certain results with acid and base 
catalysts, including the observation [8] that the 



mutarotation of tetramethylglucose is negligible in 
either cresol or pyridine, but rapid in a mixture of 
the two solvents 6 Lowry [10] advanced a mechanism 
that involves the addition of a proton at one point in 
the sugar molecule and simultaneous elimination of a 
proton at another point. The process may be 
represented in the following manner : 7 



HCOH 



HC— OHB 



HC = 



(HO.HC) n O-f-B + HA^(HO.HC) n OHA<=KHO.HC) n -f-HB + 



— C 

I 



— C 



I 
— COH 

I 



+ A- 



(3) 



This mechanism would give rise to a third-order 
term in the rate expression [11], but as pointed out 
by Swain [12], this is not detectible because of the 
relative magnitude of the terms, and hence the rate 
constant on the basis of this mechanism can be 
represented within the experimental error by eq 1. 
The simplicity of the mechanism makes it most 
attractive for interpretation of the mutarotation 



reaction, expecially in amphoteric solvents. Certain 
experimental facts, however, may be explained more 
satisfactorily by separate acid- and base-catalyzed 
mechanisms, and such mechanisms have come to be 
quite generally accepted. It seems probable that 
they act concomitantly with Lowry's mechanism, 
and that the extent to which each mechanism is 
effective is determined by the experimental conditions. 



HCOH HCOH HC=OH + 

l\ *i "l\ k 2 \ 

R O + HA <=± R OHA <=± R 

1/ *-l 1/ k- 2 | 



Acid-catalyzed mechanisms 

HOCH 



-C 

I 



— C 



— COHA- 



I 



U 



-A- 

I 



HOCH 

h |\ 

R OHA <=> R O + HA 



k-4 



-A- 



(4a) 



HCNH 2 

l\ *. 
R O + HA <=> 

I 



HCNHj 

l\ h 

R OHA <z» 

1/ fc-: 

— o 

I 



all open-chain forms 

HC=NH+ 



I 
R 

I 

-COHA- 

I 



h 

fcT 3 



H 2 NCH 

l\ k< 

R OHA <± 

1/ fc-4 

— c 

I 



H 2 NCH 

l\ 
R O+HA 

1/ 
— C 

I 



(4b) 



all open-chain forms 



HCOH 

l\ 
R O + B- 



HCO- 

l\ 
► R O + HB; 

1/ 
— C 
I 



rHco- 

l\ 

R OHB 

1/ 
-C 

I 



HC=0 

i R 

I 
-COH 

I 



■OCH 

l\ 
R OHB 



-A- 

i 



OCH HOCH 

l\ l\ 

I R + HB^± R O + B- 

1/ 1/ 

— c -c 



("Base-catalyzed" mechanism of [14]) 



(5) 



6 Similarly, Butler, Smith, and Stacey [9] have recently observed that the 
a- and /3-tetraacetates of iV-phenylglucosylamine are extremely stable in dry 
pyridine, but mutarotate to an equilibrium value on addition of a drop of water. 



7 Although the mechanism of eq 2 was represented in a generic fashion with 
respect to HA and B, in eq 3 and succeeding mechanisms charges have been 
indicated for clarity. Nevertheless, the mechanisms may be regarded as generic. 
Thus, if B in eq 6a is a neutral substance, then HB becomes HB+. 



136 



HCOH 


HCOHB- HC 


l\ 


fci | \ k 2 


R O + B- 


<± R <=> R 


1/ 


k-i \/ fc- 2 J 


-C 


— C -C 



Base-catalyzed 



II u 



IT 

all open-chain forms 



in 


echanisms 










BHOCH 




HOCH 




h 


l\ 


hi 


l\ 


B 


t~» 


R 


— > 


R O + B- 




fc- 3 


1/ 


fc-4 


i 






— C 

1 





(6a) 



HCNH 2 HCNH 2 B" HC= 

I \ fc I \ hi 

R O + B- <=± R O <=± R 

I / fc-i 1/ A;-2 

— C -C 

I ! 



— co- 

I 
IT 

all open-chain forms 



:NH BH 2 NCH H 2 NCH 

k 3 \\ h \\ 

+ HB <=> R O ^± R O + B- 

k- z | / k-i | / 

-C -C 

I I 



(6b) 



HCOH 

l\ 
R O+HB; 

-V 



HCOH 

l\ 
► R OH++B- 

I 



rHCOHB- 

l\ 
R OH"* 

1/ 
-C 

I 



HC=0 

I 
: R + 

I 
— COH 

I 



HB; 



BHOCH 

I N 
R 

I, 
— C 

I 



OH^ 



("Acid-catalyzed" mechanism of [14]) 



HOCH 



l\ 
R OH++B- 

1/ 
-C 

I 



HOCH 

l\ 
<=> R O + HB 

1/ 
-C 



I 



(7) 



In cq 4 and 6 there are proposed acid- and base- 
catalyzed mechanisms that we believe represent 
reasonable courses for the mutarotation of the sugars 
and the glycosylamines. These mechanisms differ 
in certain respects from the mechanisms commonly 
accepted. 

The acid-catalyzed mechanism presented here for 
the mutarotation of the sugars (eq 4a) begins with 
the addition of an acid catalyst to the ring oxygen. 
This is followed by rupture of the ring through an 
electron shift in which a lone electron pair of the 
glycosyl oxygen facilitates release* of electrons to the 
ring oxygen. The product quickly establishes equi- 
librium with all species of the open-chain modifica- 
tion. Cyclization can then proceed through com- 
bination of carbon 1 with the oxygen of either carbon 
4 or carbon 5. This gives rise to the alpha and beta 
furanoses and pyranoses in equilibrium proportions. 

The acid-catalyzed mutarotation of the glycosyl- 
amines is formulated in similar manner, as shown 
in eq 4b. In this case, a shift of a lone electron pair 
of tin* amino nitrogen results in the formation of the 
imonium ion — C=NH 2 + . Reaction of carbon 1 of 

the imonium ion with the oxygen of carbon 4 or 
carbon 5 produces the various ring isomers. A 
mechanism of this type will account for the striking 
sensitivity of the mutarotation of the glycosylamines 
to acid catalysts. 

A base-catalyzed mechanism for the mutarotation 
of the sugars may be represented as shown in eq 6a. 
Combination of the base with the hydrogen of the 
glycosidic hydroxyl gives an activated complex that 
decomposes with separation of the conjugate base 
and rupture of the ring. By the reversible addition 



of a proton, the product establishes equilibrium with 
the open-chain form of the sugar. Restoration of 
asymmetry on carbon 1 through ring formation, 
facilitated by combination of the conjugate base 
with the carbonyl oxygen establishes equilibrium of 
the cyclic modifications. The glycosylamine, how- 
ever, does not react to a great extent by a base- 
catalyzed mechanism (eq 6b) analogous to that of 
the free sugar because there is less tendency on the 
part of the amino nitrogen to release a proton. 

The above mechanisms differ from those generally 
accepted primarily in regard to the species of the 
sugar undergoing reaction. According to Hammett 
([13, p. 337]) " . . . the only reasonable mechanisms 
are the following: for the acid catalysis the mobile 
and reversible addition of a proton to the ether oxy- 
gen followed by a rate-determining reaction with a 
base . . . for the base catalysis the mobile and 
reversible removal of a proton, followed by a rate- 
determining reaction with an acid." These mecha- 
nisms, originally advanced by Fredenhagen and Bon- 
hoeffer [14] and based in part on the work of Bon- 
hoeffer and Reitz [15], have been extended to the 
glycosylamines by Howard, Kenner, Lythgoe, and 
Todd [16]. The mechanisms of [14] are outlined in 
eq 5 and 7, in which the steps contained in brackets 
have been postulated by us. Comparison shows that 
our acid-catalyzed mechanism of eq 4a resembles the 
base-catalyzed mechanism of eq 5, and that our base- 
catalyzed mechanism of eq 6a resembles the prior 
acid-catalyzed mechanism of eq 7. It will be seen 
that the two mechanisms classified here under acid- 
catalysis (eq 4a and 5), operate on different species 
of the sugar; this is likewise true of the two classified 
under base-catalysis (eq 6a and 7). The proportions 



137 



of the species are determined by the hydrogen ion 
concentration and the base strength of the carbohy- 
drate, and hence the relative importance of the sev- 
eral mechanisms in any solution depends on the sub- 
stance and the pH. 

Presumably all of the mechanisms, in addition to 
Lowry's, are applicable to the sugars, with eq 5 and 
6a favored in alkaline solution and 4a and 7 in acid 
solution. It appears that eq 4b would be the main 
course for the primary and secondary amines and 
indeed the only one that is applicable to tertiary 
amines. The mutarotation of tertiary glycosyl- 
amines was observed by Kuhn and Birkhofer [17], who 
concluded that the reaction involves the formation 
of a substituted ammonium ion followed by elimina- 
tion of a proton, to give the imonium ion — C = NH 2 + . 

I 
Although eq 4b involves the same imonium ion, we 
believe that it is produced directly from the amine 
rather than from the corresponding ammonium ion. 
It should be pointed out that the mechanism of eq 
4a and 4b requires only an acid catalyst and is at 
variance with the currently held concept that both 
an acid and a base catalyst are necessary for the 
mutarotation reaction. As applied to the alkyl 
glycosides, a mechanism analogous to eq 4a will ac- 
count adequately for acid-catalyzed interconversion 
in nonaqueous solvents, a reaction for which mecha- 
nisms analogous to those of eq 5, 6a, and 7 are inade- 
quate because the aglycone group does not permit 
attack by a base catalyst. 

Apparently the acid- and base-catalyzed reactions 
of the sugars take place side by side, and the over-all 
reaction can be regarded as the sum of the parallel 
and competing reactions, the rates of which depend 
on the acid and base catalysts present. It is of 
interest to examine the newly postulated mechanisms 
to ascertain whether they will account for the effect 
of acid and base catalysts on the rate of mutarota- 
tion. For this purpose, the mechanisms of eq 4a 
and 6a can be formulated in the following manner: 



Acid-catalyzed system: 



* + HA<=>X, 
fc-i 



k 2 

Xi<Z±X2 

k- 2 



X2<=^X3 

k- z 



ki 
X 3 <=>/3-hHA 

fc-4 



Base-catalyzed system: 



a + B^Xj 

fc-i 

k 2 

k- 2 

h 
X2 + HB^=±X3 

fe- 3 

k A 
k-4 

In these expressions, HA represents an acid 
catalyst, B a base catalyst, a and /3 the modifica- 
tions of the sugar, and X 1? X 2 , and X 3 reaction 
intermediates. The rate constants and the inter- 
mediates are not the same for the two systems. 
Application of the Christiansen equations ([13, p. 
107]) to the acid- and base-catalyzed systems for the 
sugars, and combination of rate constants, lead to 
the following expressions, respectively for the rates 
of reaction: 



d[a]__d[(3]_ 



k A [RA](k[a]-k'[P]), 



dt dt 
for catalysis by an acid, and 

for catalysis by a base. At equilibrium, 
d[a] 



dt 



-0. 



Thus k\a\— k'\/3] = Q in both cases, and k\k' is fixed 
by the equilibrium ratio. By adjusting the arbitrary 
constants k A and k B , the same value in the two 
equations is assigned to k and likewise to k' , so that 
(k[a]—k f \0j) becomes a common factor for all catalysts. 
The over-all rate then becomes 

_^M = (± ki[}i A t ] + ±k J [B } ])(k[a]-k'm, (8) 

for i acid catalysts andj base catalysts. If the only 
catalysts are oxonium ion, hydroxyl ion, and water 
eq 8 reduces to 



d[a] 



= (iH 2 o + fcH[H+]+t O H[0H-])(fc[a]-fc / [i8]), (9) 



dt 



where &h 2 o includes both acid and base catalysis by 
the water molecule. This equation, when integrated 
and expressed in terms of optical rotation by Hud- 
son's method [18] 8 becomes 



s At the time that Hudson presented his classical development of the kinetics 
of mutarotation, it was thought that the equilibrium of lactose was established 
between a ring form and an open-chain form (referred to in his article as lactone 
and hydrate, respectively). 



138 



7 /« r -^ JL "=(A-„. ) o+fcH[H+]+^oH[OH ~])(k+k'). 

By reducing bo common logarithms and combining 

constants, 



* log £>_£»=& H 2 o+A:H[H + ]+fcoH[OH-]=fc OT . (10) 



This is the familiar expression that Hudson found 
to represent the mutarotation of glucose in acid, 
neutral, and alkaline solutions [6]. This derivation 
shows that the postulated acid- and base-catalyzed 
reactions are in accord with the previously known, 
experimentally determined relationship correlating 
the rate constant for the mutarotation of the sugars 
with the catalysts present. 

The Christiansen equations may be similarly 
applied to the mutarotation of the glycosylamines 
(eq 4a and 4b), and the summation then takes the 
form: 



d\a\ 

dt 



(±k i [BA i ]+±k } [B,]) Ob[otaJ-f OhhJ). 



(11) 



The expression calls for general acid and base 
catalysis. It has been found experimentally that 
the mutarotation rate is increased by the presence 
of ammonium, phosphate, and carbonate buffers. 
However, for systems in which the only catalysts are 
water, oxonium ion, hydroxyl ion, and ammonion ion, 
eq 11, when treated in the same manner as eq 8, 
reduces to 



1 i r - 



^h 2 o + A:h[H+] + 



fro H [OH-]+£ N „ 4 [NH 4 +]=^. (12) 



In the case of the glycosylamines, there is theo- 
retically a complicating factor. In the derivation of 
the expression for the rate of mutarotation of the 
sugars, it is possible to assume that a direct relation- 
ship exists between the concentrations of the alpha 
and beta isomers and the optical rotation regardless 
of the presence of either more or less acid. This 
assumption is valid because only a minute part of 
the sugar exists under any given conditions in other 
than the normal forms. Because of the basic properties 
of the amino group, however, part of both the alpha 
and the beta glycosylamine probably exist in solution 
as the corresponding substituted ammonium ion. 
The proportion of the free amine in aqueous solutions 
is determined by the equilibrium: 



CH-NH 2 CH-NH+ 

i \ fa T\ 

R + H+0^± R + H 2 0. 

I / fc-i I / 

-c -c 

I I 

(a or 0) (a or fi) 



(13) 



When this reaction is taken into account, it can 
be shown by application of the mass law that 



1 i r — r« 



&h 2 o+*-h[H+] +AWOH-] + *nh 4 [NH+] 



1+K[H+ 



rCr, 



(14) 



where K is the equilibrium constant for eq 13. Ap- 
plication of eq 14 to the data of table 2 gives a series 
of simultaneous equations that show that in the pH 
range where the mutarotation is sufficiently slow 
to be measurable, the effect of the term K [H+] is less 
than the experimental error. Consequently, the 
term can be neglected and eq 12 can be applied to 
the mutarotation, of L-arabinosylamine. 

From the data of table 2, numerical values were 
obtained for the catalytic coefficients of eq 12. At 
low concentrations of oxonium and ammonium ion, 
the quantity &h 2 o+&oh[OH~] approaches k m , and 
each term must be less than the lowest observed rate 
(0.0004 at pH 12); hence &h 2 o and k on are less than 
4X10 -4 and 4X 10~ 2 , respectively. From experi- 
ments 2 and 3, & N h 4 was found to be 10~ 2 . Inasmuch 
as &h 2 o and &oh[OH~] are negligibly small in com- 
parison with the values of k m in experiments 4, 5, and 
(>, values of /"n were calculated on the assumption that 
& m =& H [H + ] + 10- 2 [NH 4 + ]. The average value, of k H 
so obtained is 6.9X 10 6 . The mutarotation constants 
of table 2, corrected for the catalytic effect of am- 
monium ion, are given in the last column of the table 
and are shown by the dotted curve of figure 3. 9 A 
comparison of the numerical values of the catalytic 
coefficients for L-arabinosylamine with those for 
D-glucose is given in table 3. It may be seen that the 
principal catalyst for the mutarotation of D-glucose 
is the hydroxyl ion, whereas for the mutarotation of 
L-arabinosylamine it is the oxonium ion. The 
striking difference arises primarily from the fact that 
the glycosyl oxygen has a greater tendency than the 
nitrogen to release a proton (eq 6a and 6b) and less 
tendency than the nitrogen to share a lone electron 
pair (eq 4a and 4b). 



Table 3. 



Catalytic coefficients of mutarotation reactions at 
20° C 



Catalyst 


Symbol 


L-Arabinosyl- 
amine 


D-Olucose a 


H 2 

H+ 

OH- 


&H 2 

fcOH 
fcNH 4 


<4X1(T 4 
6.9X10 6 
4X10- 2 
1X10 -2 


2.6X1 o -4 
3.6XKT 1 

8X10 3 
1.2X10 -3 



Data from [19]. 



e In a preliminary publication [2], an expression was given to represent em- 
pirically the rate of mutarotation in experiments 1, 7, 8, and 9. No attempt was 
made to evaluate the general acid and base catalysis of the buffers used. 



139 









A 




! 
1 

1 
Q 






.10 






/ \ 




' 










/ \ 




1 


o 
o 

g .08 










1 

1 MUTAROTATION 

1 

1 






! 




1 






z 
5 




i 


• 










RATE CONSTANTS 

b b 
-p> CD 












I 










•\ 


1 

\ 
1 

4 
\ 






/ HYDROLYSIS * 




\ 

\ 
\ 
\ 




.02 









X 










UP 


-1 

\ 
\ 

\ 
\ 




2 4 6 


8 10 12 








D 


H 









Figure 3. Rate constants for mutarotation and hydrolijsis of 
L-arabinosylamine. 

IV. Hydrolysis Reaction 

It has already been mentioned that the rate of 
hydrolysis of L-arabinosylamine was calculated from 
the increase in optical rotation that follows the 
initial decrease. The rate of hydrolysis obtained 
from the data of table 1 is shown in figure 3 as a 
function of pH. The results bring out clearly the 
striking sensitivity of L-arabinosylamine to hydrolysis 
within a limited pH range. 10 

A satisfactory mechanism for this reaction must 
account for the ease of hydrolysis and the restriction 
of the reaction to a limited pH range. It follows 
from the well-known stability of amines to acid 
hydrolysis n that the hydrolytic reaction is not 
characteristic of the C— NH 2 linkage but must be due 
to the glycosyl structure. It was mentioned in 
connection with the acid-catalyzed mutarotation 
that the glycosylamine is able to form the imonium 
ion, -C=NH 2 + , by rupture of the ring following the 

addition of a proton to the ring oxygen. A mecha- 
nism for hydrolysis including this substance as an 
intermediate is given in the next column. 12 With this 

io The reaction-rate-pH curve is similar in shape to many of the enzyme 
activity-pH curves, among which may be cited those of urease [20], 0-glueosidase 
[21], a-amylase [22], and cellulase [23]. It seems possible that the sensitivity of 
certain enzyme systems to a change in pH may also arise from equilibria compa- 
rable to those cited here. 

n The stability of the C— N linkage of amines is illustrated by the fact that 
glucosamine is obtained by treatment of chitin with concentrated hydrochloric 
acid [24]. 

12 The glycosylamines are somewhat analogous to the dialkylaminomethyl 
alkyl ethers, which have been studied by Stewart and Bradley [251 These 
authors found that the compounds are rapidly hydrolyzed by acid and explained 
the ease of hydrolysis by a mechanism involving the formation of an iminium ion 
(imonium ion) and condensation of this with hydroxyl ion. Our mechanism 
includes analogous steps. 



structure (D), addition can take place by attachment 
of a nucleophilic group to the carbon, and a shift of 
electrons. If the nucleophilic group is hydroxyl, 
addition yields an intermediate aldehyde ammonia 
(F). This grouping is known to be unstable and 
gives up ammonia readily. Thus the aldehydic 
properties of the glycosylamine account for the 
facility with which ammonia is eliminated. 

Proposed mechanism -for the hydrolysis of glycosyl- 
amines: 



CHNHo 

l\ 
R O 

1/ 
-C 

I 
.(B) 



(HfO)fc! 
(HaO)*-, 



(H 2 0)IU 2 

HC=NH+ 

I 
R 

I 
— COH 

I 

(D) 

H 

I 
HO— C-NH 2 

I 
R 

I 
— COH 

I 
(F) 

fc_ 5T j(H 2 0) 
(H+0)W-fc 5 

H 

I 
-0— C-NH 2 

I 
R 

I 
— C— OH 

I 



(H 2 0)A:3 
(H+0)/c 3 



CH-NHJ 

l\ 
R O 

-V 

I 

(C) 



HC=NH 

I 
R 

i 
-COH 

i 



(E) 



(HfO)*« 

(NH 3 )(H 2 0)fc- 



Sugar 

(X) 



(G) 

It will be shown next that the proposed reaction 
mechanism will account for the fact that hydrolysis 
is rapid only in a limited pH range. Presumably 
the first step in the hydrolysis is the formation of the 
imonium ion (D) by acid catalysis; this is accom- 
panied by a side reaction to form the glycosylammo- 
nium ion (C). The second step is the addition of a 
hydroxyl ion to D, to yield the aldehyde ammonia 
(F) ; it is also accompanied by a side reaction forming 



140 



the imine (E). The last part of the process covers 
the elimination of ammonia from the aldehyde am- 
monia. It is represented as proceeding through the 
ionization of F (reaction 5), followed by an acid- 
catalyzed rearrangement of the anion (G) to yield the 
aldehyde form of the sugar, and ultimately the equi- 
librium mixture. The ammonia is converted to am- 
monium ion, thus driving the reaction to completion. 
In the derivation of the rate constant as a function of 
the oxonium ion concentration, it is postulated that 
in the pH range in which hydrolysis occurs, equilibria 
for all reactions except 6 are established quickly and 
maintained throughout the reaction. It is also as- 
sumed that within this range reaction 6 is irreversible. 
Inasmuch as equilibrium conditions are postulated 
for all reactions except 6, the rate of hydrolysis is 
represented by 



d(A-X) 
dt 



= [G][H+]t fll 



(15) 



where A is the sum of all components and X is the 
sugar formed by hydrolysis. From mass law it fol- 
ows that 

[F][H 2 0]fr 5 =[G][H+]K 5 

fD][OH-]/fc 4 =[F]*_ 4 

[E][H+]*_ 3 =[D][H 2 0]& 3 

[B][H+]£ 2 =[D][H 2 0]A:_ 2 

[C][H 2 0]*_ 1 =[B][H + ]ifc 1 

The equations may be solved for B, C, D, E, and F 
in terms of G, and the values substituted in the ex- 
pression A=B+C+D+E+F+G+X to obtain 



A* _ 5 At _ aK . 



( A -v=w(mry; 



[YL+]k.,k_Jc^k, 



[H + }k- 5 k. 



,k 4 k 2 ' [HtOUOR-lkJcJcak. 

4 . /L_K/t_4#t3 



\H t O][OH-]k s k t T [OB.-]k i k i k. 



[H+]fc. 



[H 2 0] k 6 



') 



This, when simplified by combining constants, and 
expressing the results in terms of new constants 
becomes 



[G] = 



(A-X)[OH- 



~[OK-] 2 +k'[OH.-)+k" 



If the derived value for [G] is substituted in eq 15, 
it becomes 

d(A-X)_ (A-X)[OH-] 2 [H+]& 6 



dt [OB--] 2 +k'[OR-]+k" 

which may be simplified by multiplying numerator 



and denominator by [H + ], and combining constants 
to obtain 



d(A-X)_ 



(A-X) 



dt 



K'+K"[H+]+K'"[OH-] 



After integration, evaluation of the integration con- 
stant and conversion to common logarithms, the 
equation becomes 



1, A 1 

t 10g (A-X)~K 1 +K 2 [H+]+K 3 [OH-] 



=k 



hydrol* 



where & hydrol is the hydrolysis rate constant. Inas- 
much as the change in optical rotation during the 
hydrolysis period is a measure of the change in 
(A— X), the expression in terms of optical rotation 
becomes 



1 



t 2 t\ 



log 



r t —r a 



,=k 



hydrol* 



■r„ Kx+K 2 IH+]+K,[OH-]- 

(16) 
By use of the data of table 4, nine simultaneous 
equations corresponding to eq 16 were set up. Com- 
bination of these in groups of 4, 2, and 3, and solution 
by the method of averages gave values of 7.9, 6.1 X10 4 
and 4.2 X10 8 , respectively, for Ki, K 2 , and K 3 . Thus, 



"*hvdrol — 



1 



7.9 + 6.1 X10 4 [H+] +4. 2X10 8 [OH~] 

A piot of this expression is shown by the solid curve of 
figure 3; the experimental results are in good agree- 
ment with the curve. Thus, the postulated mecha- 
nism and the equation developed from kinetic con- 
siderations account for the fact that hydrolysis is 
rapid only in a limited pH range. 

Table 4. Summary of data for the hydrolysis reaction at 20° C 



Experiment 


PH 


[H+] 


[OH-] 


fc a hydrol 


10 


7.2 
6.8 
6.2 
5.9 

5.0 
4.6 

4.0 
3.7 
3.1 


0.63X10-7 

1.6X1G- 7 

6.3X10-7 

13X10-7 

1.0X10-5 
2.5X10- 5 

1.0XK)- 4 
2. OX 10-" 

7. 9X10-* 


16X10-8 
6. 3X10- 8 
1.6X10-8 
0.8X10-8 

1X10-9 
0.4X10- 9 

10x10-11 

5X10-H 
1.3X10-11 


0.019 
.029 

.055 
.068 

.108 
.108 

.068 
.047 
.021 


11__. 


12 ... 


13-.- 


14 


15 


16 . 


17 


18 





a fchydroi = 1/ Cf 2 — 1\) log (r«j— ra,)/(r/ 2 — r m ), where r ti is the first measurement 
in the period used for evaluating the rate of hydrolysis. 

In the preceding discussion, it was necessary to 
omit consideration of the individual reaction con- 
stants and of catalysis by acids and bases other 
than oxonium and hydroxyl ions. 13 Furthermore, 
it was not possible to consider the several ring 
isomers of the system separately. In the mechanism 
on page 140, intramolecular reaction of carbon 1 

ia Inasmuch as equilibrium conditions prevail during the hydrolysis, general 
catalysis in the formation of the imonium ion does not affect the rate of hydroly- 
sis, although it accelerates the rate of mutarotation. 



141 



of D with the hydroxyl of carbon 4 would give the 
alpha and beta furanosylamines; reaction with the 
hydroxyl of carbon 5 would give the alpha and 
beta pyranosylamines. Hence B and C symbolize 
all cyclic modifications of the glycosyl amine, and the 
corresponding cation, respectively, and the constants 
of reactions 1 and 2 represent composite rates. 
Similarly, F includes both stereomeric modifications 
of carbon 1, and the constants of reaction 5 represent 
composite rates. 



V. Evidence for the Formation of an Un- 
known Product, Possibly the Digly- 
cosylamine 

In some cases the mutarotation and hydrolysis 
appear to be complicated further by a side reaction 
that takes place during the early part of the reaction 
period and varies in importance with the experi- 
mental conditions. Evidence of this complicating 
factor may be seen from the data of table 5. It 
will be observed that when an excess of acid is 
added slowly to L-arabinosyl amine, the optical rota- 
tion drops to a lower point than when the acid is 
added quickly. The product having the lower rota- 
tion appears to be hydrolyzed more rapidly than the 
other, as the optical rotation of the solution rises 
more rapidly and ultimately exceeds that of the 
solution containing less of the byproduct. It seems 
probable that the complication is caused by the 
formation of a diarabinosylamine. Evidence of this 
reaction is also found in experiments 19 and 20 of 
table 1. In these, the initial observed rotation is 
lower than usual, and the early change in optical 
rotation is too rapid to be due entirely to the hydroly- 
sis of L-arabinosylamine. As the reaction, after the 
first few minutes, appears to follow the first-order 
course, the calculations of the rate were made with- 
out the use of the early anomalous values. In experi- 
ment 18 of table 1, however, although the initial 
rotation is likewise low, the entire mutarotation 
follows the first-order course. The side reaction, 
which under other conditions may become of major 
importance, is being studied further. 



Table 5. Reaction of ~L-arabinosylamine with acid (evidence 
for a side reaction) 

0.25 g~of L-arabinosylamine dissolved in sufficient 2.5 N HC1 to give a volume 

of 25 ml 



Time after begin- 
ning the addi- 
tion of acid 


Optical rota- 
tion after rapid 
addition of 
acid 


Optical rota- 
tion after drop- 
wise addition 
of acid during 
2.5 minutes 


Minutes 

3.8 

7.0 
15.0 
30.0 
60.0 

Change in [al 2 D ° 
in 1 hr. 


r i20 

69.0 
69.3 
69.7 
69.9 
70.0 

1.0 


MS 

67.2 
68.2 
69.4 
70.5 
71.0 

3.8 





VI. Role of the Imonium Ion in Other 
Reactions of Glycosylamines 

It is believed that the imonium ion, considered to 
be an intermediate in the hydrolysis reaction, ac- 
counts not only for the formation of the diglycosyl- 
amine but for several other reactions. Thus the addi- 
tion of a nucleophilic substance, as for instance a 
cyanide ion, would take place most readily through 
the imonium ion, and the concentration of this 
cation would largely determine the reaction rate. 
The imonium ion also plays a part in reactions of the 
Amadori type; mechanisms for some of these were 
presented in a prior publication [26]. 

VII. Experimental Details 
1. L-Arabinosylamine 

Dry ammonia was passed slowly, at room temper- 
ature, into 100 ml of methanol containing 1 g of 
ammonium chloride and 20 g of L-arabinose in suspen- 
sion. Addition of ammonia was discontinued when 
the sugar had entirely gone into solution. Ether 
was then added to the point of incipient turbidity, 
and the mixture w^as allowed to stand in the refriger- 
ator until crystallization of L-arabinosylamine oc- 
curred (3 weeks). The crystals were collected on a 
filter, washed with methanol, and air-dried; the 
yield was 15 g. Recrystallization was conducted by 
dissolving the crude arabinosylamine in 30 ml of a 
1:10 mixture of concentrated ammonia and water. 
The solution was filtered and diluted with 60 ml of 
methanol saturated with ammonia, and then with 
150 ml of absolute ethanol. The resulting crystals 
were separated after 2 hours, washed with methanol 
saturated with ammonia, and dried at room tempera- 
ture in a vacuum desiccator containing phosphoric 
anhydride. The recrystallized product weighed 
10.8 g, and melted at 124° C. At a concentration of 
2 g/100 ml in C0 2 -free water, [a]^ =+86.3° (5 min); 
83.8° (60 min): 78.6° (7 hr) ; 80.5° (54 hr). De 
Bruyn and Van Leent [2], who first prepared the 
compound, reported mp 124° C and [a^+SS 
(H 2 0, c=10). 

Figure 1 shows the titration curve obtained when 
0.1 iV HC1 was added portionwise to 75 ml of a 
solution containing 0.1493 g of L-arabinosylamine. 
Sufficient time was allowed after each addition of 
acid for the pH of the solution to reach equilibrium. 
At the end of the experiment, the optical rotation 
corresponded to that of L-arabinose, and after the 
ammonium chloride and excess acid were removed 
by treatment with ion-exchange resins, 0.14 g of 
crystalline arabinose was obtained from the solution. 
This is a nearly quantitative yield. 

2. Tetraacetyl-L- Arabinosylamine 

A mixture of 4 g of finely powdered L-arabinosyl- 
amine, 40 ml of pyridine and 16 ml of acetic anhydride, 
in a flask equipped with a mechanical stirrer, was 



142 



kept in .in ice bath and stirred until solution was 
complete (4 hrs). After 18 hours at 0° C, the solu- 
tion was poured into a mixture of ice and water, and 
the product was extracted with chloroform. Evap- 
oration of the chloroform gave a crystalline residue 
that was recrystallized from hot ethanol. The 
yield was 3.65 g. After several recrystallizations, 
tetraacetyl-L-arabinosvlamine melted at 177° to 
178° C; [a] 2 D = + 89.6°\CHCh, e=.1.6). 

Analysis: Calculated for C 13 H 19 8 N: C, 49.21; H, 
6.04; N, 4.41; COCH 3 , 54.26. Found: C, 49.3; 
H, 6.1; N, 4.4; GO-CHa, 52.4. 

3. 7V-Acetyl-L-Arabinosylamine 

Six grams of tetraacetyl-L-arabinosylamine was 
dissolved in 100 ml of anhydrous methanol, and 10 
ml of 0.1 N barium methylate in methanol was 
added. The alcohol was removed by distillation 
under reduced pressure, with exclusion of moisture. 
The residue was dissolved in water and treated 
successively with a cation and an anion exchange 
resin to remove barium salts. The resulting solu- 
tion, after evaporation, gave 3.6 g of crystalline 
product. To recrystallize, the material was dissolved 
in 6 ml of hot water, and the solution was treated 
with a decolorizing carbon. After filtration, the 
solution, combined with 2 ml of washings, was diluted 
with 2 volumes of methanol, and the mixture was 
allowed to stand for several hours at 0° C. The 
crystalline iV-acetylarabinosylamine was separated 
and washed with methanol. Melting point, 222° to 
224° C; fag? =+69.1° (H 2 0, c=4). The constants 
were unchanged after further crystallization. 

Analysis: Calculated for C 7 H 13 5 N: C, 43.97; H, 
6.85; N, 7.33. Found: C, 44.1; H, 6.8; N, 7.4. 

4. Periodate Oxidation of 7V-Acetyl-L- 
Arabinosylamine 

It was found that the optical rotation of a solu- 
tion that was 0.1 M with respect to iV-acetyl-L-arab- 
inosylamine and 0.25 M with respect to sodium 
metaperiodate (NaI0 4 ) quickly changed from a dex- 
tro to a levo direction and attained a maximum 
([ a ]™ = — 49 ) in 14 minutes. After this time, there 
was a very gradual decrease in levo-rotation. At 
intervals, 10-ml aliquots of the solution were with- 
drawn, and each sample was treated with 10 ml of 
a saturated solution of sodium bicarbonate, 25 ml 
of 0.1 N sodium arsenite, and 1 ml of a 20-percent 
solution of potassium iodide. The excess arsenite 
was then titrated with a 0.1 iV solution of iodine. 
The values in table 6 show that two moles of period- 
ate were consumed per mole of glycosylamine in 10 
minutes, and that further consumption of periodate 
was negligible. These results are in harmony with 
a pyranoside structure for A^-acetyl-L-arabinosyla- 
mine. The change in optical rotation that takes 
place during the oxidation is also given in table 6. 
The levorotatory product, presumably 



HNAc 
H 2 C— CHO OHC— CH 

I o^ 

is characteristic of all iV-acetyl-pentapyranosyla- 
mines having the same configuration for carbon 1 as 
that of 7V-acetyl-L-arabinosyl amine. 

Table 6. Oxidation of N-acetyl-ij-arabinosylamine with 
sodium metaperiodate 



Reaction 
time 


Optical ro- 
tation a 


Moles peri- 
odate per 
mole amine 


Minutes 
2.5 
3.4 
5.0 
5.3 
7.3 

10.0 
10.7 
14.2 
30 
60 
120 


r I 20 
[a[D 

+19.6 

-6.9 

~-30.'9 _ 
-42.8 

-47.1" 
-49.0 

~-44.~8~ 


~T92~~ 
1.99 

"~2~q5"" 

~~2."6i " 



a Based on weight of iV-acetyl-L-arabinosylamine. 

5. Mutarotation and Hydrolysis Measurements 

The course of the mutarotation and hydrolysis 
reactions was followed by measurements of optical 
rotation on solutions of L-arabinosylamine containing 
known amounts of acids and bases. In each experi- 
ment, 0.5 g of crystalline L-arabinosylamine was 
placed in a 25-ml volumetric flask. Sufficient acid, 
base, or buffer solution at 20° C was added rapidly to 
give a volume of 25 ml, and measurements were made 
of the optical rotation in a 4-dm tube. Time was 
measured from the moment the solvent was added 
to the crystals. At the end of the experiment, and in 
some cases at intermediate points, the pH of the 
reaction mixture was measured. As 1 mole of 
ammonia is liberated during the hydrolysis, it was 
not possible to maintain the hydrogen ion concentra- 
tion constant, and in each case the pH of the solution 
increased somewhat as the reaction proceeded. For 
purpose of comparison, the values obtained at the 
end of the experiment were used in the calculations. 

In solutions more acid than pH 7.8, the initial 
decrease in optical rotation that characterizes the 
mutarotation was too rapid to detect. In the pH 
range 7.8 to 12, velocity constants for the mutarota- 
tion were calculated bv the formula 



1 



Tt x —Tn 



H — &1 / 



where r t and r t2 are the specific rotations at times 
#i and t 2 , and r m , the minimum optical rotation, is 
assumed to be the equilibrium rotation in the 
absence of hydrolysis. In some experiments the 
minimum rotation was not reached because the 



916783—51- 



143 



reaction either was too slow, or was complicated by 
competitive hydrolysis. In these cases, a value of 
+ 69°, the minimum rotation found for uncompli- 
cated cases, was used for r m . 

It is noteworthy that the equilibrium rotation 
obtained in highly alkaline solution in the presence of 
2.5 N ammonium chloride (experiment 3, table 1), is 
appreciably lower than that found at the same pH 
in the presence of only 0.1 N ammonium chloride 
(experiment 2). The cause of the difference is not 
known, but it may be due to the side reaction 
mentioned on page 142. 

The rate of hydrolysis was calculated from the rise 
in optical rotation that takes place after the primary 
mutarotation reaction is complete. In the calcula- 
tions it is assumed that the second change in rotation 
is due solely to the hydrolysis reaction and that the 
optical rotations of both the arabinosylamine and the 
product, L-arabinose, remain constant throughout 
the period chosen for the rate measurement. 

The values for the hydrolysis rate constant given 
in table 1 were calculated from the formula 



1 



U-t 






r t 2 -r« 



where r t and r t2 are the rotations at times t x and t 2 , 
respectively, and r<x> is the specific rotation of the 
fully hydrolyzed material. 

VIII. Summary 

It has been found that L-arabinosylamine under- 
goes a mutarotation reaction that is followed or 
accompanied by hydrolysis of the amino group. The 
mutarotation takes place at measurable rates 
between pH 7.8 and 12, but it is too rapid to observe 
in solutions of higher oxonium ion concentrations. 
Mechanisms are proposed for the mutarotations of 
the glycosylamines and of the sugars. The mutaro- 
tation of the glycosylamines, in marked contrast to 
that of the sugars, is not appreciably catalyzed by 
hydroxyl ion, but is strongly catalyzed by oxonium 
ion. Apparently mutarotation and hydrolysis of the 
glycosylamines take place through an intermediate 
imonium ion. 

The rate of hydrolysis of L-arabinosylamine is 
unusually sensitive to the acidity of the medium, is 
greatest at pH 5, and is negligible in both alkaline 
and highly acid solutions. A mechanism is proposed 
to account for these facts and by application of mass 
law to the reaction scheme and evaluation of the 
constants, the following expression is obtained for 
the rate of hydrolysis: 



1 



^hydrol" 



~7.9+6.1X10 4 [H+]+4.2X10 8 [OH-] 



To provide information concerning the structure 
of L-arabinosylamine, the following new compounds 



were prepared: Tetraacetyl-L-arabinosylamine 
(C 13 H 19 8 N), mp 177° to 178° C, [«]*> = +89. 6° 
(CHC1 3 . c=1.6); iV-acetyl-L-arabinosylamine (C 7 - 
H 13 5 N), mp 222° to 224° C, [a]g , = + 69.1° (H 2 0, 
c=4). The latter substance was found to react 
quickly with 2 moles of sodium periodate, as required 
for an iV-acetylpentosylamine having the pyranose 
structure. 



The authors gratefully acknowledge the assistance 
of Nancy B. Holt in various phases of the work, and 
of Rolf A. Paulson, who analyzed the compounds 
reported. They also express their appreciation to 
T. D. Stewart of the University of California for 
kindly and helpful criticism of the reaction mecha- 



nisms. 



IX. References 



[1] H. L. Frush and H. S. Isbell, J. Research NBS 41, 609 

(1948) RP1943. 
[2] H. S. Isbell and H. L. Frush, J. Am. Chem. Soc. 72, 

1043 (1950). 
[3] C. A. Lobry de Bruyn and F. H. Van Leent, Rec. trav. 

chim. Pays-Bas. 14, 129 (1895). 
[4] P. Brigl and H. Keppler, Z. physiol. Chem. 180, 38 

(1929). 
[5] C. Niemann and J. T. Hays, J. Am. Chem. Soc. 62, 

2960 (1940). 
[61 C. S. Hudson, J. Am. Chem. Soc. 29, 1571 (1907). 
[7] J. N. Bronsted and E. A. Guggenheim, J. Am. Chem. 

Soc. 49, 2554 (1927). 
[8] T. M. Lowry and I. J. Faulkner, J. Chem. Soc. 127, 

2883 (1925). 
[9] K. Butler, F. Smith, and M. Stacey, J. Chem. Soc. 
1949 3371 

[10] T. M. Lowry, J. Chem. Soc. 127, 1371 (1925). 

[11] G. E. K. Branch and M. Calvin, Theory of organic 
chemistry, p. 404 (Prentice-Hall, Inc., New York, 
N. Y., 1944). 

[12] C. G. Swain, Abstracts of papers, p. 47P (116th Meeting, 
Am. Chem. Soc, Atlantic City, N. J., Sept. 21, 1949). 

[13] L. P. Hammett, Physical organic chemistry (McGraw- 
Hill Book Co., Inc., New York, N. Y., 1940). 

[14] H. Fredenhagen and K. F. Bonhoeffer, Z. physik. Chem. 
[A] 181, 392 (1938). 

[15] K. F. Bonhoeffer and O. Reitz, Z. physik. Chem. [A] 179, 
135 (1937). 

[16] G. A. Howard, G. W. Kenner, B. Lythgoe, and A. R. 
Todd, J. Chem. Soc. 1946, 855. 

[17] R. Kuhn and L. Birkofer, Ber. deut. chem. Ges. 71, 
1535 (1938). 

[18] C. S. Hudson, Z. physik. Chem. 44, 487 (1903). 

[19] T. M. Lowry and G. F. Smith, Rapports sur les hydrates 
de carbone, p. 116, 10th Conference of the Inter- 
national Union of Chemistry (Liege, 1930). 

[20] S. F. Howell and J. B. Sumner, J. Biol. Chem. 104, 619 
(1934). 

[21] K. Hill, Ber. Verhandl. sachs Akad. Wiss. Leipzig, 
Math. phys. Klasse. 86, 115 (1934). See also, W. W. 
Pigman and R. M. Goepp, Jr., Carbohydrate chem- 
istry, p. 490 (Academic Press, New York, N. Y., 1948). 

[22] E. Ohlsson, Z. physiol. Chem. 189, 17 (1930). 

[23] G. A. Greathouse, Textile Research J. 20, 227 (1950). 

[24] C. S. Hudson and J. K. Dale, J. Am. Chem. Soc. 38, 
1431 (1916). 

[25] T. D. Stewart and W. E. Bradlev, J. Am. Chem. Soc. 
54,4172 (1932). 

[26] H. S. Isbell, Ann. Rev. Biochem. 12, 206 (1943). 

Washington, July 25, 1950 



144